1393 research outputs found
Sort by
Comparing Unidirectional, Bidirectional, and Word2vec Models for Discovering Vulnerabilities in Compiled Lifted Code
Get PDFRansomware and other forms of malware cause significant financial and operational damage to organizations by exploiting long-standing and often difficult-to-detect software vulnerabilities. To detect vulnerabilities such as buffer overflows in compiled code, this research investigates the application of unidirectional transformer-based embeddings, specifically GPT-2. Using a dataset of LLVM functions, we trained a GPT-2 model to generate embeddings, which were subsequently used to build LSTM neural networks to differentiate between vulnerable and non-vulnerable code. Our study reveals that embeddings from the GPT-2 model significantly outperform those from bidirectional models of BERT and RoBERTa, achieving an accuracy of 92.5\% and an F1-score of 89.7\%. LSTM neural networks were developed with both frozen and unfrozen embedding model layers. The model with the highest performance was achieved when the embedding layers were unfrozen. Further, the research finds that, in exploring the impact of different optimizers within this domain, the SGD optimizer demonstrates superior performance over Adam. Overall, these findings reveal important insights into the potential of unidirectional transformer-based approaches in enhancing cybersecurity defenses
Acting Quickly: Determining Urgency in Phishing Messages
Get PDFIn the modern cyberscape, phishing serves as a useful means for attackers to gain access to their victims. Creating the best phishing message is a sophisticated art form, with one common tactic involving asking the victim to follow through on an action with great haste. Through the utilization of sentiment analysis, or weighing how much of a certain emotion is conveyed in text form, classifying this sense of urgency becomes relatively trivial. The research done in this project attempts to create a working machine learning model that can categorize the urgency value of a body of text. The data gathered from the model is used to visualize how urgent large data sets can be, and potentially can be used to determine the connection between phishing likelihood and urgency scores
AI-Powered Operations: Navigating Ethics, Automation, and Strategic Innovation in the Digital Era
Get PDFArtificial Intelligence (AI) is redefining Operations Management (OM) by transforming how organizations plan, execute, and optimize their core processes. This TREO Talk presents a comprehensive synthesis of AI’s evolution and impact in OM, underscoring its role as a strategic enabler of agility, resilience, and innovation in digitally driven enterprises. Leveraging a Systematic Literature Review (SLR) guided by PRISMA methodology across nine scholarly databases, this research constructs a validated framework that captures the multidimensional role of AI in transforming operational strategies, capabilities, and outcomes across industries. AI\u27s influence spans predictive analytics, robotic process automation, supply chain orchestration, logistics, and real-time quality assurance. For instance, Amazon’s AI-infused inventory systems dynamically anticipate demand shifts, while UPS’s ORION platform uses AI to optimize delivery networks in real time. Siemens’ smart factories deploy AI-driven computer vision and predictive maintenance to reduce downtime and improve throughput. These implementations exemplify how AI facilitates precision, scalability, and responsiveness - empowering organizations to operate more intelligently, minimizing waste, and delivering higher value to stakeholders in increasingly volatile markets. To guide effective integration, this work introduces a multidimensional Adoption Framework comprising: (1) Antecedents—technological readiness, strategic alignment, and data infrastructure; (2) Implementation—structured planning, integration, and performance evaluation; (3) Challenges—ranging from ethical risks and algorithmic opacity to workforce displacement and compliance mandates; and (4) Outcomes including enhanced operational visibility, cost optimization, and accelerated decision cycles. The framework serves as a strategic blueprint for organizations aiming to scale AI adoption effectively, ensuring alignment between technological advancement and business value creation while navigating the inherent risks of digital transformation. The future of AI in OM is shaped by three paradigm shifts. First, AI-driven sustainability is revolutionizing energy and resource efficiency, enabling greener, more responsible operations. Second, ethical AI is emerging as an operational imperative, demanding transparency, accountability, and algorithmic fairness through explainable AI and bias audits (Rai, 2020; Sharma & Sheth, 2021). Third, human-AI collaboration is empowering hybrid intelligence, where machine learning augments human judgment to enhance creativity, contextual decision-making, and adaptability in complex operational environments. This synergy fosters a new model of collaborative operations that leverages the strengths of both humans and machines. This abstract is a call to action for the Information Systems (IS) community to lead in shaping the next frontier of intelligent operations. It offers scholars a rigorous framework for future research, provides educators with insights to inform curriculum design, and delivers practitioners a strategic roadmap for deploying ethical, sustainable, and scalable AI across value chains. Ultimately, this work positions AI not merely as a technological enhancement, but as a transformative force for building the next generation of intelligent, ethical, and resilient operational ecosystems
Optimized Face Recognition Using Reinforcement Learning and Deep Learning Feature Extraction
No full textFace recognition is highly dependent on computer vision, artificial intelligence, and biometrics. Its usage is steadily increasing with systems involving security and user authentication to only cite a few. However, accurate and robust face recognition is not easy to achieve due to varying conditions such as lighting, facial orientations, and obstructions. It also requires substantial computational resources, which is also a constraint for the deployment of face recognition on resource-constrained devices. These limitations show the need to strike a balance between accuracy, robustness, and efficiency when implementing face recognition. This paper proposes a robust and adaptive hybrid feature-extractor-based face recognition method that fuses two lightweight deep learning models: MobileNetV2 and EfficientNetB0. A support vector machine is used for accurate classification. Reinforcement learning, implemented through Q-learning, is used to dynamically optimize the contribution weights α and β for both feature extractors. The system preprocessed the input images, generated hybrid embeddings through a weighted combination of deep features, and changed those weights to achieve optimal performance. The results in both the training and testing data sets were excellent, with an accuracy of 97% and the predictions were verified by confusion matrix analysis with low processing times (27 to 58 ms per step). The main contributions of this work include: the efficient integration of hybrid embeddings for complementary feature representation, a dynamic Reinforcement learning-based weight optimization, and a robustness of the model against variations in lighting, facial orientations, and obstructions. The proposed method is promising for real applications regardless their challenges such as computational overhead and sensitivity to image quality
The Art of Campaigning: Joint Planners Working at the Intersections of Everything
Get PDFThis article examines the evolution of military campaigning from traditional large-scale combat operations to ongoing efforts that shape the global environment and achieve long-term strategic goals. It highlights the role of Combatant Command Campaign Plans (CCPs) in aligning national strategies with operational objectives and addressing modern challenges like cyber, cognitive, and information warfare
Toward extracting scattering phase shift from integrated correlation functions. III. Coupled channels
Get PDFThe formalism developed in the preceding papers that connects integrated correlation function of a trapped two-particle system to infinite volume scattering phase shift is further extended to coupled-channel systems in the present work. Using a trapped nonrelativistic two-channel system as an example, a new relation is derived that retains the same structure as in the single channel, and has explicit dependence on the phase shifts in both channels but not on the inelasticity. The relation is illustrated by a exactly solvable coupled-channel quantum mechanical model with contact interactions. It is further validated by path integral Monte Carlo simulation of a quasi-one-dimensional model that can admit general interaction potentials. In all cases, we found rapid convergence to the infinite volume limit as the trap size is increased, even at short times, making it potentially a good candidate to overcome signal-to-noise issues in Monte Carlo applications
A Teacher Apprenticeship Pathway in a Rural, Midwest State: Perspectives of Teacher Apprentices
Get PDFThe national teacher shortage makes it challenging for principals and superintendents to hire certified teachers. To address this problem, a university in a rural state in the Midwest partnered with their state’s agencies to develop a teacher apprenticeship pathway (TAP) for 78 paraprofessionals working in the state’s public, non-public, and tribal schools. The TAP provides an organized pathway to earn a teaching degree. This study reveals the perceptions of the participants after completing their first semester. The results reveal that most experiences are positive, yet they desire more communication and help with time management. The results of this study can be useful to principals and superintendents who may be partnering with stakeholders to begin a TAP
Pure Nash Equilibrium and Strong Nash Equilibrium Computation in Additive Aggregate Games
Get PDFAggregate games, first conceptualized by Nobel laureate Reinhard Selten in 1970, model the decision-making of interdependent agents where each agent’s utility depends on their own action and the aggregation of everyone’s actions. We consider computational questions on pure Nash equilibrium (PNE) and pure strong Nash equilibrium (SNE) for aggregate games. On the way, we define a new subclass of aggregate games we call additive aggregate games, which encompasses popular games like congestion games, anonymous games, Schelling games, etc. We show that PNE existence is NPcomplete for very simple cases of additive aggregate games. We devise an efficient aggregate-space algorithm for determining the existence of a PNE and computing one (if exists) for bounded aggregate space. For SNE, we show that SNE recognition is co-NPcomplete and SNE existence is Σ^_2-complete, even for simple types of additive aggregate games. For large classes of aggregate games, we provide several novel and efficient aggregate-space algorithms for recognizing an SNE and deciding the existence of an SNE. Finally, we connect our results to several well-studied subclasses of aggregate games and show how our computational schemes can shed new light into these games
Toxicity in State Sponsored Information Operations
Get PDFState-sponsored information operations (IOs) increasingly influence global discourse on social media platforms, yet their emotional and rhetorical strategies remain inadequately characterized in scientific literature. This study presents the first comprehensive analysis of toxic language deployment within such campaigns, examining 56 million posts from over 42 thousand accounts linked to 18 distinct geopolitical entities on X/Twitter. Using Google’s Perspective API, we systematically detect and quantify six categories of toxic content and analyze their distribution across national origins, linguistic structures, and engagement metrics, providing essential information regarding the underlying patterns of such operations. Our findings reveal that while toxic content constitutes only 1.53% of all posts, they are associated with disproportionately high engagement and appear to be strategically deployed in specific geopolitical contexts. Notably, toxic content originating from Russian influence operations receives significantly higher user engagement compared to influence operations from any other country in our dataset. Our code is available at https://github.com/shafin191/Toxic_IO
Exploring FinTech trends in Jordan: insights from the post-COVID-19 era
No full textPurpose
This study aims to investigate how the COVID-19 pandemic has impacted the growth of FinTech within the Jordanian context. Specifically, it examines the antecedent factors influencing users’ intentions to adopt and use FinTech services during this unprecedented period. Design/methodology/approach
A quantitative research method approach with partial least squares-structural equation modelling was used to test the research proposed model. Findings
The results of the structural path revealed that the drivers of behavioural intentions to use FinTech services are perceived usefulness, perceived self-efficacy, perceived vulnerability, subjective norms, perceived severity, perceived certainty and resistance to use and these factors explain 79.5% of the variance of behavioural intention. Research limitations/implications
This study identifies key factors for FinTech acceptance in Jordan, offering actionable insights for policymakers and service providers. Building user trust in service and technology (security, effectiveness and transparency) is crucial, alongside leveraging social influence (testimonials, endorsements) and boosting awareness/knowledge through campaigns and educational initiatives. This study identifies key factors for FinTech acceptance in Jordan, offering actionable insights for policymakers and service providers. Building user trust in service and technology (security, effectiveness and transparency) is crucial, alongside leveraging social influence (testimonials, endorsements) and boosting awareness/knowledge through campaigns and educational initiatives. This study’s generalizability is limited by its sample, drawn from three Jordanian regions and dominated by potential users. Comparing actual and potential user perceptions, alongside including users from diverse contexts (e.g. other developing countries), could address this. In addition, using qualitative methods alongside the quantitative approach and exploring indirect relationships (mediating/moderating) would provide richer insights into FinTech acceptance in Jordan. Originality/value
While prior research has explored FinTech adoption, including studies conducted in Jordan, this study makes several key contributions. Firstly, it specifically examines FinTech trends within the post-COVID-19 era in Jordan, a period marked by accelerated digital transformation and evolving user behaviours. Secondly, it extends existing models of FinTech adoption by incorporating “perceived self-efficacy, perceived vulnerability, subjective norms, perceived privacy, perceived severity, perceived value, perceived certainty and resistance to use”, in addition to the established factors of “perceived usefulness and ease of use”. This expanded model provides a more holistic understanding of the drivers and barriers to FinTech adoption in this context. Finally, this research provides empirical evidence from the Jordanian market, offering valuable insights for financial institutions and policymakers seeking to promote FinTech adoption and financial inclusion in a developing economy undergoing rapid digital change. This focus on the post-pandemic Jordanian context, combined with the extended theoretical model, offers a novel contribution to the literature